History Says Amazon Stock's 2 Worst Years of the Past 15 Ended in Net Losses, and Both Rebounds Were Huge

Amazon plans ~$220B in 2026 capital expenditures (raised from ~$200B), including incremental AWS GPU capacity with Nvidia for 2026–2028. The article argues that unlike 2014 and 2022—when operating income deteriorated and net losses followed—current fundamentals look supportive, with operating income up 43% YoY to ~$27.5B and AWS revenue up 37% YoY last quarter. Shares trade near ~$266, up ~15% in 2026, with the key risk being whether operating income continues rising as the build-out proceeds.

Analysis

The market is likely underpricing the distinction between "capex heavy" and "capex value-destructive." For AMZN, the immediate read-through is not the size of spend but whether AWS growth stays high enough to absorb depreciation and keep operating leverage positive; that is the real support for the multiple over the next 1-3 quarters. If operating income keeps compounding, the free-cash-flow drawdown is a funding choice, not a deterioration signal.

Second-order winners sit one step removed from the headline: NVDA benefits from the multi-year GPU deployment schedule, but that upside is more about order visibility than fresh surprise, so the stock may already be discounting a lot of the 2026-28 demand. The bigger incremental beneficiary could be the broader AI infrastructure stack, while the losers are anyone funding growth without visible earnings conversion. If the market starts to worry that AI hardware amortization outpaces monetization, duration-sensitive tech multiples can compress even if revenue is still accelerating.

Contrarian risk: consensus is probably too focused on negative free cash flow and not enough on ROIC, but the bear case is that AI spend has a shorter economic life than prior cloud buildouts. The key falsifier is simple: if AWS revenue growth rolls over or Amazon's operating income stalls in the next 1-2 quarters, the stock can quickly re-rate from "investment phase" to "margin squeeze."

Key Decisions for Investors

• Long AMZN on any 3-5% pullback over the next 2-6 weeks; thesis is continued AWS operating leverage. Falsifier: next quarter operating income flat to down or AWS growth materially below current run-rate.
• Relative value: long AMZN / short NVDA over 1-3 months if you want the cleaner expression of AI spend monetization. AMZN has more room for earnings revision; NVDA has the demand story but less incremental surprise.
• Use NVDA as a slower-burn beneficiary rather than a chase buy; add only on weakness, since much of the 2027-28 GPU demand is now visible and may already be in the stock.
• No action in AAPL, NFLX, GETY, or SYBT on this headline alone; the article does not change their fundamental path and forcing a trade here is low edge.